308 Professors Ayrton and Perry on Direct-Reading 



strand should be at all symmetrical in the way in which they 

 are wound on the bobbin. With some of our earlier commu- 

 tator-instruments the resistance of the contact of the springs 

 with the metallic pieces of the commutator introduced an error 

 when the ten wires were each thick and short, since then the 

 terminal resistance was a very material portion of the whole 

 resistance of any one of the ten wires. To avoid this the 

 commutators and the springs, as will be seen from the various 

 specimens of the instruments on the table, which, Darwinian- 

 like, show the history of the survival of the fittest, have assumed 

 a much larger form. 



In our direct-reading commutator-instruments the scale is 

 double, as may be seen from fig. 2 ; the upper numbers 

 corresponding with the number of volts when the commutator 

 is to parallel, the lower when to series ; while in a Direct- 

 Reading Commutator Ammeter the upper numbers correspond 

 with the number of amperes when the commutator is to series, 

 and the lower when to parallel. 



Fig. 3 shows the interior of the latest form. A A is 

 the bobbin round which the wire is coiled ; MM is the per- 

 manent controlling magnet, made in two parts to obtain 

 considerable strength ; P P are the soft-iron pole-pieces of 

 special shape, to be referred to further on ; F F are the soft- 

 iron cores, and B B is a connecting bridge-piece of charcoal- 

 iron used in our latest form of instrument for increasing the 

 strength of the deflecting field for a given current by con- 

 necting magnetically these iron cores together. This addition 

 is very important in voltmeters ; since, although in an am- 

 meter the constant of the instrument is nearly independent of 

 temperature, this is not the case in a voltmeter. The current 

 producing a given deflection in either instrument is practically 

 independent of temperature ; but the potential difference 

 necessary to be maintained at the terminals of the instrument 

 to produce the current, and which potential difference is what 

 the voltmeter is designed to measure, depends necessarily on 

 the resistance and therefore on the temperature of the coil. A 

 temperature-correction can be easily made for the portion of 

 the effect arising from the temperature of the room ; but the 

 unknown amount of heating of the coil arising from the pas- 

 sage of the current is far more difficult to allow for. Hence 

 it is generally necessary to be contented with a weaker con- 

 trolling permanent magnet for a voltmeter in order that the 

 magnetic effect necessary to produce any given deflection, and 

 consequently the heating effect which is proportional to the 

 square of this magnetic effect, shall be less in the case of a volt- 

 meter than in an ammeter. But the employment of our soft- 



